• CARLOS FRESNEDA

    London

Updated Wednesday,1November2023 - 11:35

  • Share on facebook
  • Share on twitter
  • Send by email

Comment

The first global summit on the safety of Artificial Intelligence (AI) has kicked off at Bletchley Park, 80 kilometres from London, with a new open letter on the "catastrophic risks" of the development of the technology signed by 250 experts and sponsored by Canadian scientist Yoshua Bengio and Professor Yui Zeng, China's top academic authority on the subject.
"We call on the world's governments to actively respond to the potentially catastrophic risks that advanced AI systems can cause to humanity," the open letter reads. "This responsibility does not fall just on a few shoulders, but on the strength of the global community. Our future is at stake. We must ensure that AI is developed in a way that is safe, responsible and for the betterment of all humanity."
Yoshua Bengio and fellow Canadian Geoffrey Hinton, the neural network expert who left his position at Google this year and expressed fears of the uncontrolled advance of new technologies, will be two of the most eminent voices at the AI safety summit, in contrast to Meta's chief scientist, Yann LeCun. which considers fatalistic predictions to be "exaggerated and absurd".
More than a hundred academic experts, as many representatives of technology companies, fifty countries (including Spain) and representatives of civil society and multilateral organizations converge for two intense days at the mansion in the English countryside where Alan Turing and his team cracked the code of the German Enigma machine during the Second World War and laid the foundations for computer science.
Prime Minister Rishi Sunak will serve as master of ceremonies and aims to reach an agreement at the summit to create a global panel on AI safety, comparable to the IPCC on climate change and with the capacity to assess progress and risks year after year. Sunak has also anticipated the creation of the so-called AI Security Institute in the United Kingdom, which aims to consolidate its status as the vanguard of artificial intelligence in Europe.
The first step has been taken this week, however, by the United States, with the presidential order recently signed by Joe Biden and measures such as the obligation for AI companies to share their safety tests with the government and ensure that they do not endanger national security, the economy or public health.
U.S. Vice President Kamala Harris will in fact have a special role at the summit and will simultaneously give a long-awaited speech in London. "We have a moral, ethical, and societal duty to ensure that AI advances in a way that protects the public from potential harms and ensures that everyone benefits," Harris said.

Find out more

Future.

The first global summit on the challenge of Artificial Intelligence: "We are playing with technology like a child plays with a bomb"

  • Written by: CARLOS FRESNEDA (Correspondent)London
  • Written by: ILLUSTRATION: JOSETXU L. PIÑEIRO

The first global summit on the challenge of Artificial Intelligence: "We are playing with technology like a child plays with a bomb"


The Bletchley Park summit will focus on so-called frontier Artificial Intelligence, or "machine learning models that exceed current capabilities and can perform a wide variety of tasks." Several scientists have warned, however, that focusing on the future can serve as a distraction from the "more tangible and immediate" threats that technology is already creating in the present: from disinformation to deepfakes, from disruption in elections to the impact on the labor market this decade.
"The focus of the debate should be around disinformation," warns expert Aidan Gomez, executive director of Cohere, who will contribute his experience in the technology behind chatbots to the summit. "Extremely convincing models are being created, virtually indistinguishable from human-made text and images. This is something we need to tackle urgently. We have to find a way for the public to distinguish real information from information that isn't."

  • Artificial intelligence